广东工业大学学报 ›› 2017, Vol. 34 ›› Issue (04): 78-83.doi: 10.12052/gdutxb.160079

• 综合研究 • 上一篇    下一篇

远离最差解的粒子群优化算法

刘文凯, 温洁嫦   

  1. 广东工业大学 应用数学学院, 广东 广州 510520
  • 收稿日期:2016-06-12 出版日期:2017-07-09 发布日期:2017-07-09
  • 通信作者: 温洁嫦(1964–),女,教授,主要研究方向为最优化方法与智能计算.E-mail:985028593@qq.com E-mail:985028593@qq.com
  • 作者简介:刘文凯(1990–),男,硕士研究生,主要研究方向为优化算法与机器学习.
  • 基金资助:

    广东省教育厅特色创新资助项目(2014KTSX055)

An Improved Particle Swarm Algorithm Far Away from Worst Solution

Liu Wen-kai, Wen Jie-chang   

  1. School of Applied Mathematics, Guangdong University of Technology, Guangzhou 510520, China
  • Received:2016-06-12 Online:2017-07-09 Published:2017-07-09

摘要:

针对传统粒子群优化算法容易陷入局部最优、寻优精度低及后期搜索速度慢等缺陷,提出一种参考局部最差解影响的粒子群算法.当算法搜索后期,全局最优解(global best solution,Gbest)无变化时,局部最优解(personal best solution,Pbest)等于Gbest,这时速度靠拢最优方向向量为零,粒子前进的方向只有自身惯性.而本文加入了局部最差(partial worst solution,Pworst)之后的算法使粒子的前进方向不仅受自身惯性的影响,而且可以继续的寻优,从而找到Gbest.算法采用远离全局最差解和局部最差解的思想,对粒子群优化算法的速度更新公式进行改进,并分别测试全局最差解和局部最差解对粒子群优化算法的影响.通过几个典型的测试函数仿真结果表明,改进后的算法在搜索速度、寻优精度、鲁棒性方面较粒子群算法有了显著提高,而且具有跳出局部最优的能力.

关键词: 粒子群优化算法, 全局最差, 局部最差

Abstract:

A new particle swarm optimization (PSO) algorithm is proposed in reference to worst solution of being local in standard PSO to avoid the problems of being easy to converge to a local minimum, premature convergence and low precision. In the late period of particle swarm optimization and global optimum (Gbest) when there is no change, the local optimum (Pbest) is equal to the Gbest. Then the speed to move closer to the optimal direction vector is zero, and particles go along the direction only by its own inertia. When the proposed algorithm is added to the local minimum (Pworst) algorithm, the particle's forward move is not only affected by its own inertia, but it also can continue to find the best, so as to find the global optimal. This algorithm, by applying the idea of the particles avoiding the worst local and global solution, improves the velocity updating formula based on the PSO, and tests respectively the worst global solution and local worst solution on the influence of particle swarm optimization algorithm. Several typical test function simulations show that the proposed algorithm not only has great advantages of convergence property over some other modified PSO algorithms, but is also effective in avoiding being trapped in local optimal solution.

Key words: particle swarm optimization (PSO), global worst

中图分类号: 

  • TP301

[1] KENNEDY J, EBERHART R C. Particle swarm optimization [J]. Proc IEEE Int Conf Neural Networks, 1995, 4(8): 1942-1948.
[2] SCHMITT M, WANKA R. Particle swarm optimization almost surely finds local optima[J]. Theoretical Computer Science, 2014, 561: 57-72.
[3] SHI Y, EBERHART R. Modified particle swarm optimizer [C]// IEEE International Conference on Evolutionary Computation Proceedings. Anchorage, Alaska, USA: IEEE Service Center. 1998: 69-73.
[4] 赵志刚, 黄树运, 王伟倩. 基于随机惯性权重的简化粒子群优化算法[J]. 计算机应用研究, 2014, 31(2): 361-363.ZHAO Z G, HUANG S Y, WANG W Q. Simplified particle swarm optimization algorithm based on stochastic inertia weight [J]. Application Research of Computers, 2014, 31(2): 361-363.
[5] 张选平, 杜玉平, 秦国强, 等. 一种动态改变惯性权的自适应粒子群算法[J]. 西安交通大学学报, 2005, 39(10): 1039-1042.ZHANG X P, DU Y P, QIN G Q, et al. Adaptive particle swarm algorithm with dynamically changing inertia weight [J]. Journal of Xi’an Jiaotong University, 2005, 39(10): 1039-1042.
[6] RATNAWEERA A, HALGAMUGE S K, WATSON H C. Self-organizing hierarchical particle swarm optimizer with time-varying acceleration coefficients [J]. IEEE Transactions on Evolutionary Computation, 2004, 8(3): 240-255.
[7] P N SUGANTHAN. Particle swarm optimizer with neighborhood operator [C]// Evolutionary Computation. Washington DC, US: Procc IEEE Congr Evol Comput. 1999: 1958-1962.
[8] 任凤鸣, 李丽娟. 改进的PSO算法中学习因子(c1, c2)取值的实验与分析[J]. 广东工业大学学报, 2008, 25(1): 86-89.REN F M, LI L J. Experiment and analysis of the value selection of acceleration coefficients (c1, c2) in PSO method[J]. Journal of Guangdong University of Technology, 2008, 25(1): 86-89.
[9] KENNEDY J, MENDES R. Population structure and particle swarm performance [C]// Evolutionary Computation, CEC’02. Hawaii, US: Congress on Evolutionary Computation. 2002: 1671-1676.
[10] KENNEDY J, MENDES R. Neighborhood topologies in fully informed and best-of-neighborhood particle swarms [J]. IEEE Transactions on Systems Man & Cybernetics Part C, 2006, 36(4): 515-519.
[11] 马胜蓝, 叶东毅, 杨玲玲. 一种新的粒子群拓扑设计准则[J]. 计算机工程, 2015, 41(1): 200-206.MA S L, YE D Y, YANG L L. A new design criteria of particle swarm topology[J]. Computer Engineering, 2015, 41(1): 200-206.
[12] 石松, 陈云. 层次环形拓扑结构的动态粒子群算法[J]. 计算机工程与应用, 2013, 49(8): 1-5.SHI S, CHEN Y. Dynamic particle swarm optimization algorithm with hierarchical ring topology [J]. Computer Engineering and Applications, 2013, 49(8): 1-5.
[13] 方峻, 唐普英, 任诚. 一种基于加权有向拓扑的改进粒子群算法[J]. 计算机技术与发展, 2006, 16(8): 62-65.FANG J, TANG P Y, REN C. A modified particle swarm optimization based on directional weighting Topology [J]. Computer Technology and Development, 2006, 16(8): 62-65.
[14] CHEN Y, PENG W C, JIAN M C. Particle swarm optimization with recombination and dynamic linkage discovery[J]. IEEE Transactions on Systems Man & Cybernetics Part B Cybernetics A Publication of the IEEE Systems Man & Cybernetics Society, 2007, 37(6): 1460-1470.
[15] ANDREWS P S. An investigation into mutation operators for particle swarm optimization [C]// Evolutionary Computation. Vancouver, Canada: IEEE Congress on Evolutionary Computation. 2006: 1044-1051.
[16] 张创业, 莫愿斌. 基于协同进化思想的人工鱼和粒子群混合优化算法[J]. 广西民族大学学报: 自然科学版, 2009, 15(1): 74-77.ZHANG C Y, MO Y B. AFSA and PSO hybrid algorithm based on collaborative evolution [J]. Journal of Guangxi University for Nationalities (Natural Science Edition). 2009, 15(1): 74-77.
[17] 周玉光, 曾碧, 叶林锋. 改进粒子群优化算法及其在4G网络基站选址中的应用[J]. 广东工业大学学报, 2015, 32(2): 64-68.ZHOU Y G, ZENG B, YE L F. Improved particle swarm optimization and its application in 4G network base station location [J]. Journal of Guangdong University of Technology, 2015, 32(2): 64-68.

[1] 陈璟华, 邱明晋, 唐俊杰, 田明正, 谭耿锐. 基于改进差分进化和粒子群混合算法的电力系统最优潮流计算[J]. 广东工业大学学报, 2017, 34(05): 22-28.
[2] 陈璟华, 梁丽丽, 丁林军, 周俊, 刘国祥. 计及风电场并网的机会约束规划的机组组合优化[J]. 广东工业大学学报, 2017, 34(01): 50-54,64.
[3] 周玉光, 曾碧, 叶林锋. 改进粒子群优化算法及其在4G网络-基站选址中的应用[J]. 广东工业大学学报, 2015, 32(2): 64-68.
[4] 郭经韬, 陈璟华, 周俊, 许伟龙. 基于组合混沌序列动态粒子群算法的电力系统无功优化[J]. 广东工业大学学报, 2014, 31(2): 85-89.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!